“Molecular Shakers” as Transmembrane Single‐Molecule Channels Toward 1:1 Cl⁻/K⁺ Cotransport

ABSTRACT

Cation–chloride cotransporters (CCCs) are an important family of chloride channel proteins that mediate electroneutral transport of Cl⁻ with Na⁺ and/or K⁺. In this study, we introduced ultracycles (molecular shakers—so named for their cocktail‐shaker‐like shape) designed to mimic the binding sites and functions of KCC1. These ultracycles incorporate both anion (S_Cl1, S_Cl2) and cation (S_K) binding sites within the single‐molecular backbone. The synthesis was achieved through dynamic nucleophilic aromatic substitution, starting from a rigid‐flexible macrocyclic precursor and diphenol derivatives. Planar lipid bilayer measurements demonstrated that the molecular shakers function as single‐molecule channels, mediating ion transport. By varying the structural features of the lower and larger rims of the macrocycle, as well as the glycol chains and the pH of the bulk solution, we achieved a permeability ratio up to P _Cl ⁻/P _K ⁺ = 0.83.

Shaker‐shaped ultracycles were designed to mimic the binding sites and function of human KCC1. These ultracycles function as single‐molecule channels that achieve nearly 1:1 electroneutral Cl⁻/K⁺ cotransport, and their transport behavior can be modulated by varying the structural features of the lower and larger rims of the macrocycle, the glycol chains, and the pH of the bulk solution.

1. Introduction

The search for the target of loop and thiazide diuretics led to the discovery of cation–chloride cotransporters (CCCs), a fascinating family of chloride channel proteins that mediate the coupled, electroneutral transport of Cl⁻ with Na⁺ and/or K⁺ at a 1:1 stoichiometry across the cell membrane [1]. CCCs are expressed in various tissues and organs, and play critical roles in cell volume regulation, renal salt reabsorption, intracellular chloride homeostasis, and modulation of neuronal excitability [2]. Malfunction of CCCs leads to a wide range of human diseases, including epilepsy, hearing loss, Gitelman syndrome, Bartter syndrome, and Andermann syndrome [3, 4]. As a result, CCCs have emerged as promising therapeutic targets for development of pharmacological compounds [5, 6, 7]. A cryo‐EM structure of human KCC1 reported by Liu and coworkers established the structural and functional relevance of the ion transporter [3]. KCC1 forms a dimer stabilized by the interactions between transmembrane domain (TMD) and extracellular domain (ECD). The 2.9‐Å resolution structure of KCC1, crystallized from KCl solution, revealed three ion binding sites (S_K, S_Cl1, and S_Cl2) in the ion pathway. These sites are supposed to work in a cooperative way to facilitate the stoichiometric Cl⁻/K⁺ cotransport (Figure 1a). Despite the significant physiological, pharmacological, and therapeutic importance of CCCs, their transport mechanism remains elusive. This is primarily because their electroneutral 1:1 transport stoichiometry renders the process membrane‐potential‐independent and yields no net ionic current, which precludes definitive conclusions from standard electrophysiological techniques. Moreover, in live‐cell models, this specific chloride‐transport mode is difficult to distinguish from other conductive pathways [8]. Therefore, designing synthetic channel molecules with well‐defined structures and binding sites offers a promising strategy to address these challenges and elucidate the fundamental principles of coupled ion transport. However, although numerous cation‐ or anion‐selective artificial molecular channels have been developed [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32], an artificial transporter capable of electroneutral cation–Cl⁻ cotransport has yet to be realized.

FIGURE 1.

(a) The location of potassium and chloride binding sites in the ion pathway of KCC1. (b) Design of the molecular shakers as single‐molecule channels exhibiting anion and cation binding sites.

To develop a cation–chloride cotransporter, we envisioned that multitopic ultracycles could serve as viable candidates. First, very large macrocycles (>50 ring atoms [33, 34]) could provide single‐molecule channels sufficiently long to span the bilayer membrane. Second, the multitopic design enables simultaneous cation and anion binding in a stoichiometric manner when both binding sites are incorporated within the ultracyclic backbone in close proximity. Based on this rationale, we designed shaker‐shaped ultracycles featuring tetraoxacalix[2]arene[2]triazines as anion–π binding sites (S_Cl1, S_Cl2) and pentaethylene glycol linkers as cation affinity sites (S_K) (Figure 1b). A key challenge in this design is balancing chloride and potassium transport selectivity, as chloride faces greater difficulty traversing the channel due to its larger size (1.81 Å vs. 1.38 Å) and higher hydration energy (−340 kJ/mol vs. −295 kJ/mol) compared to potassium. To address this, our strategy focuses on enhancing anion binding by incorporating hydroxy groups at the lower rim and amino acids at the larger rim of the macrocycle, facilitating cooperative anion–π and hydrogen bonding interactions. Additionally, to stabilize the channel within the bilayer, we anchored phenylalanine and serine moieties as terminal groups and introduced rigid aromatic moieties within the pentaethylene glycol linkers. Herein, we present the synthesis of several molecular shakers and their ion transport properties, specifically targeting electroneutral Cl⁻/K⁺ cotransport, as evaluated through planar lipid bilayer measurements.

2. Results and Discussion

To synthesize the target channel molecules, we conducted a one‐pot cyclization reaction between macrocyclic precursors 1 (for preparation of 1 see Scheme S1 and S2) and diphenol derivatives 2. In the presence of 8.0 equiv CsF, the reactions between 1 and 2 yielded mainly rearranged products [35], including [2+2] oxacalix[2]arene[2]triazine‐containing ultracycles 3a–c, along with [1+1] macrocyclic side products 4a–c (Schemes 1 and S3). However, under the same condition the reaction of macrocyclic precursors and 2c exclusively produced 4d, 3d could not be isolated. To synthesize 3d, we adopted an alternative approach, starting with the reaction of a linear trimer 5a and 2c in the presence of 12.0 equiv CsF, which successfully yielded the target [2+2] ultracycle (Schemes 1 and S4). The ultracyclic compounds 3a–d were then deprotected using Pd/C and H₂ as reducing agents to produce channel molecules 6a–d (Schemes 1 and S5). The synthesized macrocycles were fully characterized by spectrometric and elemental analysis (Figures S1–S8, S38–S54).

SCHEME 1.

Synthesis of ultracyclic single‐channel molecules.

With the channel molecules 6a–d in hand, we first investigated their chloride and potassium binding capabilities using ¹H NMR spectroscopy. Due to the poor solubility of the four compounds in most organic solvents, we prepared saturated solutions in CD₃CN. After filtration, the spectra of channel molecules 6a, 6c, and 6d were recorded. Chloride (as a tetrabutylammonium salt) and sodium or potassium (as tetrakis(perfluorophenyl)borate salts) were then added to the solutions, respectively. As shown in Figure 2a, the addition of Cl⁻ to 6a induced downfield shifts of the NH protons on the serine moieties, indicating the formation of hydrogen bonding between serine and chloride. Downfield shifts were also observed for the aromatic protons H^a and H^b, which can be attributed to cooperative anion–π and hydrogen bonding interactions within the electron‐deficient cavity. In line with our previous studies and a theoretical study, in addition to the ubiquitous anion–π interaction between the anion and the triazine rings of the oxacalix[2]arene[2]triazine scaffold, the low‐rim hydrogen atom can also participate in anion binding through weak hydrogen bonding [29, 36]. Similarly, when K⁺ or Na⁺ was added, the protons on the glycol linkers (H^c and H^d) exhibited downfield shifts, demonstrating the cation binding ability of the glycol chains (Figure 2b). These titrations confirmed the binding of 6a to both cation and anion. Similar ion pair binding behavior was observed for 6c and 6d. Regarding cation binding, 6a and 6c exhibited higher affinity for K⁺ than for Na⁺, whereas 6d showed the opposite trend, suggesting that the presence of a benzene ring within the glycol chain hinders potassium binding (Figures 2c–e and S9–S14). Despite these variations in cation binding, the ¹H NMR titration results indicate that the channel molecules function as ion pair receptors and exhibit potential for chloride–potassium cotransport.

FIGURE 2.

(a and b) Partial ¹H NMR spectra of 6a (saturated solution, 500 µL, CD₃CN, 800 MHz) upon addition of Cl⁻ (5.8 equiv), K⁺ (2.1 equiv), Na⁺ (2.1 equiv), representing the anion and cation binding ability. (c–e) Chemical shift changes of the protons on the anion and cation binding sites upon addition of Cl⁻, K⁺, and Na⁺, respectively. The red lines correspond to the right vertical axis, while the black line corresponds to the left vertical axis.

The transport activity was evaluated using large unilamellar vesicles (LUV) assay combined with ion‐selective electrode potentiometry (ISE). Vesicles encapsulating KCl were suspended in Na₂SO₄ solution, and chloride efflux was monitored over time using a chloride‐selective electrode. The normalized fractional effluxes revealed that 6a–d exhibited chloride transport activity compared to the DMF blank, with the activity following the order 6d > 6b > 6c > 6a (Figures 3b, S15, and S16). Since sulfate, an extremely hydrophilic anion, was used as the extravesicular buffer solution and is unlikely to be transported across the lipid bilayer, the chloride transport mediated by these compounds follows a cation/chloride synport mechanism. Hill analysis of the dose‐response curves for KCl showed that 6d (EC₅₀ = 115 ± 11 µM) and 6b (EC₅₀ = 128 ± 16 µM) are more active than 6c and 6a (EC₅₀ > 1000 µM).

FIGURE 3.

(a) Schematic diagram of the ion‐selective electrode potentiometry assay. (b) Chloride efflux mediated by 6a–d (16–32 mol% of 6a–d to phospholipid), EYPC vesicles (500 mM KCl in 10 mM HEPES, pH = 7.0) were suspended in sodium sulfate solution (333 mM Na₂SO₄ in 10 mM HEPES, pH = 7.0). Triton X‐100 was added to lyse the vesicles at 650 s. Chloride efflux was measured using a chloride‐selective electrode. (c) Evaluation of the chloride transport activity of 6d in varied concentrations by ISE assay. (d) Hill fitting of the normalized efflux at 450 s.

Confirming the macroscopic chloride transport activity of transporters, planar lipid bilayer measurements (BLM) were then conducted using 1,2‐diphytanoyl‐sn‐glycero‐3‐phosphocholine (DPhPC) lipid bilayer membrane to study the single‐molecule transport behaviors. After injection of 6a–d to the cis chamber and applying voltages ranging from −150 mV to +150 mV to symmetric KCl bath solutions (cis/trans = 1.0 M/1.0 M), typical current traces were observed (Figures 4a and S18–24). The slope of the linear fit of the current–voltage relationship (based on three repeated measurements) yielded conductances of 26.22 ± 0.90 pS for 6a, 14.36 ± 0.55 pS for 6b, 20.47 ± 0.72 pS for 6c, and 9.50 ± 0.35 pS for 6d, respectively (Figure 4b). These results indicate that 6a–d form symmetrical, single‐molecule channels within the bilayer membrane. The embedded anion and cation binding sites facilitate ion dehydration and help overcome the high energy barrier associated with transport across the hydrophobic membrane.

FIGURE 4.

(a) Current traces of 6a at +150 mV to −150 mV in symmetrical KCl solutions (cis/trans = 1.0 M/1.0 M, pH = 7.0). (b) I−V plots under symmetrical KCl solutions (cis/trans = 1.0 M/1.0 M, pH = 7.0). (c) I−V plots under asymmetrical KCl solutions (cis/trans = 1.0 M/0.5 M, pH = 7.0 or 6.5). Values are averages of three repeated experiments.

Next, we investigated the ion selectivity of 6a–d under asymmetric bath solutions. A KCl concentration gradient across the membrane (cis/trans = 1.0 M/0.5 M KCl) was established to determine the reversal potentials, which were used to calculate the permeability ratio (P _Cl ⁻/P _K ⁺) using the Goldman–Hodgkin–Katz (GHK) equation. We first examined the ion permeability ratio mediated by 6a, a compound lacking cooperative hydroxyl groups on the lower rim. At pH = 7.0, an average reversal potential of −9.56 mV was observed, corresponding to a permeability ratio of P _Cl ⁻/P _K ⁺ = 0.24. This result indicates that 6a is a K⁺‐selective transporter. Generally, K⁺ has a smaller ionic radius and lower dehydration energy than Cl⁻, making potassium selectivity easier to achieve in artificial channels. In this case, despite the incorporation of two anion binding sites (tetraoxacalix[2]arene[2]triazine [37, 38]) within the molecular backbone to enhance anion selectivity, the significant preference of 6a for K⁺ over Cl⁻ suggests that its anion binding ability is insufficient to compete with cation binding. We then turned to 6b, a lower‐rim‐hydroxylated channel molecule. Previous studies have shown that lower‐rim hydroxyl groups can enhance anion transport activity by forming cooperative hydrogen bonding and anion–π interactions with the triazine rings [29, 39]. Consistent with this, BLM measurements of 6b revealed an increased permeability ratio of P _Cl ⁻/P _K ⁺ = 0.49, confirming that the additional hydroxyl groups on the lower rim indeed improve anion selectivity. The amino acid moieties on the larger rim also influence the permeability ratio. For example, replacing serine with phenylalanine reduced chloride selectivity (P _Cl ⁻/P _K ⁺ = 0.21 for 6c). We propose that the amino acids not only anchor the channel within the phospholipid membrane through the interaction with phospholipid head groups, but also provide additional hydrogen bonding sites for anion binding. In addition to the considerations on anion binding sites, we envisioned that the modification of glycol chains would affect the chloride–potassium selectivity. Owing to the easier synthesis of phenylalanine‐terminated channel molecules, we modified 6c by inserting a lipophilic, rigid aromatic ring into the center of the glycol chain, resulting in 6d. This modification was expected to inhibit potassium binding, as suggested by ¹H NMR titration studies. As anticipated, BLM measurements showed that 6d exhibited reduced transport efficiency compared to 6c (9.50 pS vs. 20.47 pS) but an increased permeability ratio (P _Cl ⁻/P _K ⁺ = 0.63).

To further approach 1:1 Cl⁻/K⁺ cotransport, we investigated the effect of pH value on ion transport selectivity of 6d. Lowering the pH of the buffer solution was expected to enhance chloride transport selectivity by inhibiting the deprotonation of terminal carboxyl groups and strengthening chloride binding [28]. Accordingly, we measured ion transport selectivity at pH = 6.5. Surprisingly, this slight pH decrease resulted in an average permeability ratio of P _Cl ⁻/P _K ⁺ = 0.83. A further reduction to pH 6.0 yielded a comparable ratio (Figures S36 and S37). The maximum ratio of 0.83 approached near‐stoichiometric Cl⁻/K⁺ cotransport [8].

3. Conclusion

In summary, we designed and synthesized shaker‐shaped ultracycles with large dimensions and multitopic binding sites. These ultracycles, capable of spanning the bilayer membrane and forming ion transport pathways, mimic the anion–cation binding sites of the natural KCC1 protein. Through covalent modification of anion and cation binding sites and pH‐dependent regulation of anion selectivity, we achieved cation‐coupled chloride transport, replicating the functionality of natural cation–chloride cotransporters (CCCs). Given the physiological and pharmaceutical significance of CCCs, this work opens new avenues for research in both artificial and biological ion channels.

Author Contributions

Wen‐Hui Mi: data curation, investigation, validation, formal analysis, writing – original draft, methodology. Xu‐Dong Wang: methodology, data curation, validation, formal analysis, visualization, writing – original draft, writing – review and editing, supervision, conceptualization. Zhong‐Wen Chen: data curation. Yu‐Fei Ao: resources, project administration. Qi‐Qiang Wang: supervision, resources. Li‐Xia Wang: data curation, validation, formal analysis, resources. De‐Xian Wang: conceptualization, methodology, validation, supervision, formal analysis, funding acquisition, visualization, writing – original draft, writing – review and editing, resources, project administration.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73556‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.